The Agent Skills Directory

[Data Processing and Indirect Prompt Injection Surface]: The skill ingests data from external sources to define report requirements and implementation plans, which represents a potential surface for indirect prompt injection.
Ingestion Points: The agent reads content from local project files (such as .pbip, .tmdl, and .Report folders) and processes user-provided prompts during multiple clarification rounds.
Capability Inventory: The skill is capable of writing files (creating report specifications and generating PBIR structures), invoking local tools (Power BI Desktop, Node.js), and initiating network operations for cloud publishing via companion management tools.
Boundary Markers: The workflow includes a strict human-in-the-loop checkpoint, requiring the user to manually review and approve the _brief/report-spec.md file before implementation begins.
Sanitization: The skill relies on structured clarification rounds and the mandatory approval gate to mitigate the risk of processing unintended instructions embedded in project data.
[Tool Orchestration and Command Execution]: The skill acts as a high-level orchestrator for specialized Power BI automation tools.
The instructions direct the agent to invoke companion skills and MCP servers (e.g., powerbi-report-authoring, powerbi-modeling-mcp) to perform technical tasks.
These operations, including file system modifications and Fabric publishing, are standard within the Power BI report lifecycle and are delegated to specific tools to ensure consistency and control.

powerbi-report-planning